Re: [#95] Multiple Content-Lengths

On 21/09/2010, at 3:00 PM, Roy T. Fielding wrote:

> On Sep 20, 2010, at 7:14 PM, Mark Nottingham wrote:
>>> The latter changes don't work.  "it" is ambiguous, and
>>> "SHOULD NOT be used" is a new requirement that no current browser
>>> supports AFAIK.
>> 
>> Yes, but the feedback from them about making this change is positive, overall. 
> 
> ...
> 
>>> I don't think we can require anything more than it SHOULD be
>>> indicated as an error.
>> 
>> I read the discussion so far as leaning towards getting rid of that requirement.
> 
> Excuse me, but ... WTF?
> 
> So two browser developers insist that they can't possibly indicate
> an error message, but they are willing to refuse to use the response
> at all?  Implying, of course, that they will indicate an error message,
> since otherwise the user has no idea what happened to their request
> (or inline image/object/stylesheet/script).

This is why I said we need to clarify the terminology around user interaction. They read "inform the user" as printing a message (possibly in a modal dialogue box, status bar, etc.). You appear to be reading it as "show a broken image box."

It would be more helpful if you explained how you envision this requirement being fulfilled. E.g., how will a browser that receives a CSS stylesheet with two content-length headers inform the user of the problem? Does simply not displaying the stylesheet qualify?


> I'd really like to encourage that we stop asking people's opinions
> and instead ask them what they are going to implement (or already
> have implemented).  Everyone has an opinion and I'd prefer that
> thoughtful answers be provided instead of knee-jerk slogans.


I've been told (albeit privately) what they are willing to implement, and have communicated that back. I again encourage other implementers (thanks, Anne) to engage in the discussion directly. 

<chair hat>

Roy, characterising other people's opinions *or* intent as "knee-jerk slogans" doesn't move the discussion forward, it makes implementers more reluctant to engage openly on-list. Please refrain.

Thanks,



--
Mark Nottingham     http://www.mnot.net/